# VIP Daily Prompts 20–39 (from tasks.addict.best) Injected into content.addict.best · 20 prompts ## 20. screening settings ``` Recommended setttings for video screening: ------- Quality:2k, Or Native Screen Resolution (e.g., 1920x1080, Or 2560x1440) Recommended options: CRF: 18 to 21 (Use 19 CRF for 9:16 format (for duration of 45 to 55 seconds),and CRF 21 for 16:9 videos format of longer duration, more than 70 seconds ) -Tag-tag:v hvc1 is Better compatibility (especially Apple / browsers) -strong-intra-smoothing=0:aq-mode=3 -We will upload to browser web or youtube, so use yuv420 p =4:2:0 better than yuv444 in this case —Prefer no-sao=1 for screen content. ``` ## 21. visual output prompt ``` You are smart content creator and expert images, video editor, and scenarist who write attractive content and prompts for images, videos, scripts, SEO, drama, comedy manga, and comics. Other different types styles like cinematic realistic, ghibli, surrealism, and you choose best styles according to content, how it will go and show, social media accounts marketing. Give me high quality visualized output: 1-Prohibited points to be shown in the video (NO EYES, NO CIRCLES , NO FACE FEATURES, NO NOSE, NO MOUTH, NO EYES) Remember: -Vip note to consider in any prompt generation, The characters must be without faces. (Make faceless images without eyes, without mask, and no eyes circles. Remember no faces (faces may have only representative texts), and make paraphrase to this prompt and spec tasks to remember and -force flag to restrict to rules and make best output production. -NO EYES, NO CIRCLES , NO FACE FEATURES, NO NOSE, NO MOUTH, NO EYES ``` ## 22. merge to one html file ``` convert this content or merge these files into one html file with boxes, side list (not upper navigation bar) clear light and deep colors compatible with max. saturation, quality, contrast, and resolution sharpness. High speed and very rapid performance. ``` ## 23. session summary ``` Errors avoidance I want you to summarize all outputs in the session from beginning to the end, reporting actions required, how it made with code examples, and how you fix / handle issues. final results with reasons- details, how to enhance in future. -Add also in the summarized session file you will create all output files you made through this session and its full path places (full paths from user/Khaled until folder or file name) like that eample template: /Users/khaledahmedmohamed/designs-content/...). and what was the goal behind creating them and their uses. ``` ## 24. Act on server / separate -folder ``` (1) Put ALL your outputs in new unique named folder (2) Acting on server: use SSH to avoid local resource drain. (3) Transfer the output into the required folder of the user. (do not invent another folder or another root) ``` ## 25. local copy / download ``` Transfere (download from server to mac or transfere from local mac , to keep one version at mac) transfer your folders into this workspace folder, and add unique term name of folders to your folders name to be specific and unique: /Users/khaledahmedmohamed/designs-content/vip/x content/videos and description/... ``` ## 26. video simulator ``` A-For the videos present in this workspace try to identify them, see (visualize) them then tell me how to create videos like them (How to animate images ,edit images style, colors, motion, effects, transition, to generate like these videos exactly? what are types of images to generate to create such videos (Make full description for each video, focus on style, design, colors, and effects ,not content meaning) -Add each video style-design-effects-editing description in separate file .md with unique name and all files output, collect them in one folder. B-After you identify each video give it name to its style (name based on video style and title, and make for each video file with description and give the file same name of video title + .md at end of file description of prompts and design, style, effects, editing to generate same style by Ai-agent) Add prompts, pro-tips, free tools and open-sources repos from github (not heavy run, but moderate or light with power features) to help cli Ai-agent to create such videos. -Add each video style-design-animation description in separate file .md with unique name and all files output, collect them in one folder. ``` ## 27. review github apps ``` Review these github repo apps priortize them from best powerful unique features to least one, and give better, easier a or lighter alternative for each one if present: ``` ## 28. comparison / evaluation / merge ``` -compare different stacks here -compare between them in: -performance stability and speed -running resources consuming -afford high traffic users. - better and enhanced features -valid output with high performance - Easy-moderate difficulty to build without errors and high stability and valid results -Recommended edits or merging tips to generate best valid practical working stack (we have vps to host so avoid paid options, and avoid heavy running consuming resources tools) ------------------------- Merge: ``` ## 29. screen videos ``` Can you take screen recording video to different sections of the app as you are the admin try app features to show uses and benefits, and for sections which need user input data, you add user data input (like simulation, for sections needed user data input to show features of the app, and its valuable uses) So, make new videos screen recording, consdering previous notes after finishing, screen another type of videos in same way but for 9:16 so you will screen record videos for two types 16:9 9:16 and duration of each video must be from 45 seconds to 55 seconds do not make 26 or 35 seconds make from 45 to 55 seconds duration for 9:16 for 16:9 you can make from 45 seconds to 75 seconds —— screen more other sections, click all boxes until end, do not leave any section or button ot box without clicking, act and scroll and click in slow manner. navigate all parts, and click on boxes, show expanded content. —— For sections which need user input data, you add user data input (like simulation, for sections needed user data input to show features of the app, and its valuable uses) --- Avoid screening search of no results in the app, so search for present data you ensure it is in the app, and screen searched results data. --- boxes clicking and showing details: when you search for drug or procedure, click on the box to show full uses until end of the app, as you stop at just searching and showing box from outside without clicking on it to show details of the box inside, so click on the box to expand it showing its content. ``` ## 30. pdf / images to video ``` can you animate any pdf images to create the video you have visualized, scanned and analyzed them (type name of video and its style-effects prompt): (convert pdf to images, Then make the prompt with tools and tips on images to create like video, then make video animation-effects, design, motions, colors like the style of the video you described for style and design) ``` ## 31. Videos to images: ``` First step: Download images screens from video and then convert the images into pdf format (merge them) Second step: Remove faces features and remove eyes from images in the pdf ``` ## 32. apps coding condsiderations ``` Apps coding considerations: --- deployment consideration: Check dev logs for errors/broken circuits Fix any runtime/hydration issues Implement local Storage persistent bookmarks Add print stylesheet Add image loading skeletons Mobile responsiveness fixes for new features Lint + browser verify + worklog update — What is meant by this point, and its meaning: Accessibility: focus trap for lightbox + ARIA toast Found issues: CORS warning + Fast Refresh runtime error. Let me fix both and implement all remaining features fast ``` ## 33. how to use video as reference ``` direct simple way: animate like this video exactly, and create like this anime exactly in full reversal steps to get same results copy -------------- Steps to reverse: 1- Using these images (Create complementary images, and make proper edits to allow for final similar results of the animation video ( name or link) 2- from begining of creating images, scenes animations: scenes animations include images to video prompt animation and in-between scene pure video creating prompt (not necessary all prompts to be images to videos, but you are smart to generate images then convert to videos for complex scenes or scenes that need multi-step (specially for characters shown scenes) and ready prompts to create video directly (for ensured easy output video) that help to be merged with images that converted to video to make similar output of this link or this animation video ``` ## 34. SEO writer ``` English content: You must be as SEO content creator, as you write the audio script for the video Seo-content / voice over with description of the content that support SEO for each platform overgrowth, spreading, reach, and virality. Make all your outputs (videos and .md file in separate folder). ------ Arabic SEO - version: -------------------- can you make new .md version for each app arabic version (arabic mixed with english terms for voice over scripts for Arabic audience. As you are SEO content creator, as you write the audio script for the video content with description of the content that support SEO for each platform overgrowth, spreading, reach, and virality. Make all your outputs (videos and .md file in separate folder) ``` ## 35. Artifact ``` Convert artifact to .md; full path report at server + local Mac copy + concise valuable summary. CMD: Can you convert artifact to .md file and tell me file name and full file folder path at server make copy at local mac also, and tell me everything about the concise valuable summary, names and folder paths (full file folder) from user until file end the path format must be like (shape format example, not necessary use it) this: /Users/khaledahmedmohamed/designs-content/vip/x content/... ``` ## 36. PDF Reducing Size ``` for these pdf, or pptx ot ppt present convert to reduced size pdf: Second low-RAM copy of PDF hub; keep originals. CMD: create second version folder of this pdf folders hub but in lower reduced size ram, and try to keep high quality with low size space RAM (the task goal: keep the original and create second copy of it but in very low size of all the pdf) ``` ## 37. Prompt of Orchestrator ``` === Orchestrator — Multi-Agent / Multi-Model Credits === You plan, assign specialist models, review like a human expert creator, and merge outputs to save credits and maintain efficiency. CMD: Use orchestration and manage other models (you make plan, direct them to give best output for each one master in and excel in, you review, think in logic like human expert creator/editor, then collect all outputs, you can order agy or cli-Ai agent to help you in editing, or make reversal engineer to excelent output to help you in fast better results and effective team work. The idea is to mentain working efficiency, use best features of each agent and model in proper way, you can also, copy excellent master output or use it as references to generate its like design in expert way. Also this help you to distribute credits use of images and videos content on different models api keys to save and give also best output). to help you in doing this: (muapi- prodia infrence bash tool shell)- agy-grok-gocode with openrouter free visual models like nometron-omni model or others) but save credits and make best use of them (make all your out put in new folder). ``` ## 38. Negative-prompt ``` Remember: -Vip note to consider in any prompt generation, The characters must be without faces. (Make faceless images without eyes, without mask, and no eyes circles. Remember no faces (faces may have only representative texts), and make paraphrase to this prompt and spec tasks to remember. -NO EYES, NO CIRCLES, NO FACE FEATURES, NO NOSE, NO MOUTH. ``` ## 39. Prompt to expand episode ``` EPISODE EXPANSION & ANTI-HALLUCINATION RULES: - Expand single episode with scene prompts to animate images correctly without errors. - Force AI-agent to follow rules without hallucination; precise output. - No blur in video itself or text; blur only outside video text and outside frame (9:16). - Show app feature uses and real product examples inside episodes. - Expert tips for clean motion and dynamic captions. 2. RETENTION & VIRAL HOOK MATH (2026): - Hook Timing: TikTok ~1–1.3s, Reels ~1.5–2.1s. Keep 60%+ past 3s for FYP/Reels/Shorts push. - Negative / Mistake Framing: ("you're doing this wrong") beats positive framing by ~1.3–1.8× on TikTok. - Feature Demos: Show "Old Slow Way vs App Way" side-by-side or in rapid sequence. 💡 PRO TIPS: • Audio & Captions: Keep captions word-by-word timed to speech, centered vertically, no blur. • Motion: Use smooth Ken Burns pan/zoom on stills (VFR constant 30fps) with audio padded to match video (apad + -shortest). • Hook Execution: Lead with a mistake-frame visual in the first 1.2s, transition to the app UI solution by second 3. ``` ### 40. Apply effects while screening ``` **Part 1 — Applying Effects by AI-Agent While Making Screen Recordings of the App** Yes, an AI-agent can (and should) apply effects intelligently during or immediately after screen recording. Here is the practical way to do it: ### Recommended Workflow for AI-Agent + Screen Recording 1. **Record Clean First** - The agent should always capture a clean, high-resolution screen recording of the app (no filters, no effects, no overlays yet). - Prefer 1080×1920 or higher native recording so there is quality headroom. 2. **Immediate Post-Processing by the Agent** After the raw recording is finished, the agent should automatically run a processing pipeline that includes: - Force 9:16 crop/scale if needed - Apply the correct color grade for that specific app (see Part 2) - Add subtle Ken Burns or slow digital zoom on important screens - Insert kinetic captions timed to the voiceover - Add light pattern interrupts (zoom punches, highlight circles, arrows) on key UI elements - Apply micro-SFX (soft whoosh when switching screens, soft pop on important numbers) - Enforce safe zones so text and critical UI never sit under platform buttons 3. **Best Effects to Apply Automatically on App Screen Recordings** - **Clean High-Key or soft neutral grade** (most important for medical/fitness apps) - Very light vignette only if the screen feels flat - Slow Ken Burns or gentle punch-in on key results/calculations - Kinetic captions (highest impact) - Highlight circles or soft glow on important buttons/numbers - Minimal film grain (only if the content needs a slightly less “clinical” feel) 4. **What the Agent Should Avoid on Live App Screens** - Heavy grain, strong light leaks, chromatic aberration, or aggressive teal-orange grades (these reduce readability of medical numbers and UI) - Fast chaotic zooming that makes text unreadable - Any effect that distorts colors of clinical data **Practical Agent Instruction Example:** > “After finishing the screen recording, apply Clean High-Key grade, add subtle 1.1–1.15× Ken Burns on each major screen, burn kinetic captions, add soft highlight circles on key calculated numbers, and export in 9:16 1080×1920.” This approach keeps the app interface crisp and trustworthy while still giving the video the motion and retention tools needed for social media. ```